tv Google Facebook Twitter Testify Before Senate Intelligence CSPAN November 1, 2017 9:30am-12:32pm EDT
to most on this committee we have come to expect this. we have spent countless hours working through the threats that exist to this country and around the world and it's sad we have come to the point where nothing can happen that surprises us. it's the responsibility of this committee to work hand in hand to help keep america safe by providing the tools that they need. it is in spreading disinformation and discord during the 2016 elections. this was an opportunity for each of you to tell your respective
stories and if necessary correct the record it will be discussing russia and russia's activities. i'm hoping to provide with how foreign actors used their platforms to circulate lives and agitate unrest during last year's elections. i'm also hopeful you'll share with us what you're companies are doing to make it harder for foreign actors to use your platforms automatted accounts and falsified stories to reduce sentiments in the united states. very clearly this vulnerability is to an unacceptable risk and
you have the responsibility to reduce that vulnerability. i want to use this to push back on some narratives that have been sprung up around this subject. a lot of folks have tried to reduce this entire conversation to one premise. foreign actors to help elect the united states president. i'm here to tell you, this story does not simplify that easily. it is shortsighted and dangerous to selectively focus on one piece of information and think that some how tells the whole story. we have heard from the media how a series of russian linked facebook ads were aimed at michigan and wisconsin during the lead-up to last year's presidential election. some of those target groups in two states. the narrative here is that they
were directly influenced the elections outcome. what you weren't -- what you haven't heard is that almost five times more ads were targeted at the state of maryland than of wisconsin. maryland which is targeted by 262 ads in comparison to wisconsin's 55 ads and maryland was not up for grabs. it was a state they carried by 26%. or that 35 of the 55 ads targeted at wisconsin ran prior to the wisconsin primary before there was an identified republican candidate and moreover that not one of those 55 ads mentioned president donald trump by name or that the key election state of pennsylvania had fewer ads targeted than washington d.c. where 87% voted for hillary
clinton or that the three most heavily targeted states in america, maryland, missouri and new york were all determined by at least 18 point margin and two of them won by hillary clinton. one point the media has gotten correct is more of these ads ran in 2015 than 2016. g again, before president trump was identified as a republican candidate. some of them surrounding ads on hot-button issues purchased by russian actors is missing. to add some detail here where the immediate what failed to do it and put the $100,000 into a frame of reference. the total ads spent for the state of wisconsin was $1,979 with all but $54 being spent before the primary. again, before the emergence of a
republican candidate. the average ad was $823. pennsylvania is $23. to believe the narrative that they studied our process, assessed what states would be critical to the election result then snuck and invested all of $300 to execute their plan in pennsylvania. $300. more than five times as much money was spent on advertising in california, a state that hasn't voted republican and presidential elections in 1988. even with the benefit of numbers and what can be calculated and measured this is an incredibly complex story. we can look at the number of ads purchased and draw conclusions about diversity. we can look at the pages that they directed people towards an the number of tweets and retweets in the manipulated
search results and drawing inferences about the intent of the information operation. what we cannot do is calculate the meddling for can we assume that it must be the explanation for an election outcome that many didn't expect. i understand the urge to make this story simple. it's human nature to make complex manageable fine explanations and interpret things in ways that form to your conclusions. that's biassed. it doesn't prove intent or even motive. it shows no ads ran there after the election. this subject is complicated. there is a whole new vocabulary that comes with this stuff. impressions are different than views. views are different than clicks. there is one thing i'm certain
of, and it's this. given the come pleplexity of wh have seen, if anyone tells you they have it all figured out they are kidding themselves. we can't afford to kid ourselves about what happens and continues to happen today. that complexity i'll note is exactly why we depend on you for expert incite and reliable information. 60% of the u.s. population uses facebook. a foreign power using that platform to influence how americans see and think about about one another is as much a policy issue as much as a concern. it is effective but not overly burdensome demands faith and partnership. just recently on the basis of a more complete and sophisticated analysis the original estimate
that 10 million americans were exposed on facebook was incleesed to 126 million. that tells me that your companies are just beginning to come grips with the scale and the depth of the problem. that's encouraging but know t s this, we do better when you do better. i would urge you to deep that in mind and work with us proactively to find the right solution to a very constant and complaining challenge. i'll take a moment here to stress that this hearing, what this hearing is and is not about. this isn't about relitigating the 2016 u.s. presidential election. this isn't about who won or who lost. this is about national security. this is about corporate responsibility, and this is about the dlib rative and
multi-facetted manipulation by agents of of a hostile foreign power. i'll say it again, agents of a hostile foreign power using our own social media platform and conducted an information operation intended to divide our society along issues like race, immigration and second amendment rights. what's even more galling is that to tear us apart they are using social media platforms americans invented that define an open and democratic society. while it is shocking to think that foreign actors use the social networks and communications mediums that is so central to our lives today in an effort to interfere what is even more troubling is the likelihood is these are being used to spread lives and drive
americans apart. your three companies have developed platforms that have tremendous reach and tremendous influence. it is enabled by the enormous amount of data on users and their active tis. the american people need to understand how russia used that information and what you're going to protect them. your actions need to catch up to your responsibilities. we have a lot to get to this morning. i will stop here and i want to thank each of our witnesses today. i turn to the vice chairman for any comments you might have. >> thank you. let me express our concern and thoughts about the tragedy yesterday in new york. i'll get right at it. in the age of social media you can't afford to waste too much time or too many characters in getting the point across. russian operatives are attempted
to infiltrate and manipulate american social media to hijack the national conversation and to make americans angry, to set us against ourselves and at their most basic democracy. they did it during the 2016 presidential campaign. they are still doing it right now. in many ways the threat is not new. russians have been conducting information warfare for decades. what is knew is the advent of social media tools with the power to magnify propaganda and fake news on a scale that was unimaginable back in the days of the berlin wall. today's tools in many ways seem almost purpose built for russian
disofgs techniques. russia's play book is simple. it works for this. they set up thousands of fake accounts, groups and pages across a wide aray. they pop late content on facebook, twitter, you tube, read it, linked in and many other platforms. each of these fake accounts spend literally developing real people to follow and like their content, boosted by tools like paid ads and automatted bots. most real-life followers have no idea they are caught up in these webs. these networks are later utilized to push an array of disinformation including stolen e-mails, state-lead propaganda,
fake news and devicive content. >> the goal is pretty simple, to get this so called news and to suddenly push the americans in the direction the kremlin wants to go, someone who deeply respects the tech industry and was involved in that industry for more than 20 years. it has taken me quite a bit of time. i'm still learning to truly understand the nature of this threat. even i struggle to keep up with the language and mechanics. the difference in how they generate likes, tweets and shares and how all of these players and actions are combined into an online eco system. what is clear however is that
this play book offering a tremendous bang for the disinformation buck. with just a small amount of money adversaries use hackers to steal and weaponize data, fake accounts to build networks, bots to drive traffic and ads to target new audiences. they can wreak halvoc on our online discourse. it is a pretty good return on investment. so where do we go from here? i believe it will take all of us to use some of the platform companies, the united states government and the american people to deal with this new and evolving threat. the social media and innovative tools each of you have developed have changed our world for the better. you transformed everything from
shopping for groceries to growing small businesses. russia's actions are further exposing the dark underbelly of the eco system you have will b replicated by other adversaries. both nation states and terrorists that wish to do harm to democracies around the globe. this is not a unique american phenomenon. as such, each of you today needs to commit more resources to identifying bad actors and when possible preventing them from abusing our social media ecosystem. thanks in part to pressure from this committee, each company has uncovered, i believe, only some of the evidence of the ways exploited their platforms during the 2016 election. for facebook, much of the attention has been focused on the paid ads that russian tolls targeted to americans.
however, these ads are just the tip of a very large iceberg. the real story is the amount of misinformation in devisive content and pushed for free on russians, that was spread widely on news feeds of tens of millions of americans. according to the data facebook provided, 120 russian back pages built a network of over 3.3 million people. from these now suspended pages, 80,000 organic unpaid posts reached an estimated 126 million real people. more than a a third of the population. this is an astonishing reach from just one group in st. petersburg. i doubt the so-called internet research agency in st. petersburg represents the only
russian trolls out there. facebook has more work to do to see how deep this goes, including into the reach that we have just found in the last 48 hours of information you provided of ira-backed instagram posts. which if we take for for example, 80,000s posts from trolls on facebook, 120,000 pieces of content on instagram and we don't even have the data on how much that content reached. the anonymity provided by twitter and the speed by which it shares news makes this an ideal tool to spread disinformation. according to one study during the 2016 campaign, junk news actually outperformed real news in some battleground states leading up to election day. another study found that bots generated 1 out of every 5 political messages posted on
twitter over the entire presidential campaign. people who testified before this committee have estimated that up to 15% of active twitter bts are potentially 45 million plus accounts are fake or automated. despite evidence of significant incursion and outreach from researchers, twitter has to date a small piece of that activity. in the last few days the numbers have gone from 200 accounts over 2700 accounts. the russian efforts and false
stories and unen substantiated rumors were elevated on google search during the recent pass shootings in las vegas. meanwhile, youtube has become the go-to platform. google has now uncovered 1100 videos associated with this russian campaign. much more of your content was likely spread is through other platforms. but it's not just the platforms that need to do more. the united states government has thus far proven incapable of meeting this 21st century challenge. unfortunately, i believe this effort is suffering in part because of lack of leadership at the top. we have a president who remains unwilling to acknowledge the threat that russia poses.
and should address this very real threat pose bid russian propaganda. i believe that congress, too, must do more. we need to recognize that current law was not built to address these threats. on what i believe is the most light touch legislative approach, which i hope all my colleagues on this panel will review. the honest ads act is a national security bill intended to protect our elections from the foreign interference we all want to avoid. finally, but perhaps most importantly, the american people also need to be aware of what is happening to our news feeds. we all need to take a more discerning approach to what we are reading and sharing and who we're connecting with online. we need to recognize the person at the other end of that facebook or twitter argument may not be a real person at all.
the fact is that this russian weapon has already proved its success and cost effectiveness. we can be assured that other adversaries, including foreign intelligence operatives and potentially terrorist organizations have read this playbook and are already taking action. it's why we collectively must act. to our witnesses today, i hope you will detail what we saw in the last election and most importantly tell us what steps you will undertake for us to get ready for the next one. we welcome your participation and encourage your commitment to addressing this shared responsibility. thank you, mr. chairman. >> thank you, senator. i'd like to notify members we will have seven-minute rounds today. if i could ask you to please stand and raise your right hand. >> do you swear to tell the
truth, the whole truth and nothing but the truth? >> please be seated. we're going to recognize you and then mr. walker. >> chairman, vice chairman warren ir and distinguished members of the committee, thank you for this opportunity to appear before you today. my name is collin stretch, and i have served as the general counsel of facebook. we appreciate this committee's hard work to investigate russian interference in the 2016 election. at facebook our mission is to create technology that gives people the power to build community and bring the world closer together. we are proud that each of you uses facebook to connect with jr. constituents and e we understand the people you
represent expect authentic experiences when they come to our platform to share and connect. we also believe that we have an important role to play in the democratic process and a responsibility to protect it on our platforms. that's why we take what's happened on facebook so seriously. the foreign interference we saw during the 2016 election is reprehensible. that foreign actors hiding behind fake accounts abused our platform and other internet services to try to sew division and discord and to try to undermine our election process is directly contrary to our values and everything we stand for. these foreign inging a tors set drive people apart. in our investigation, which continues to this day, we have found that these actors used fake accounts to place ads on
facebook and instagram that reached millions of americans over a two-year period. and those ads were used to promote pages, which in turn, posted more content. people shared these posts, spreading them still further. many of these ads and posts are inflammato inflammatory, some are offensive. we know that much of this content is particularly hurtful to members of the facebook community that engaged with this content believing it was authentic. people should believe content is awe thebtic and should not have to worry they are being exploit ed in a cynical effort to pray on painful fault lines in our society in order to inflame discourse in this country. an aggregate, the ads we are here to discuss were a small fraction of the overall content on facebook, but any amount is too much. all of these accounts and pages violated our policies and we
removed them. going forward, we are making significant investments. we are hiring more ad reviewers, doubling or more our security engineering efforts, putting in place tighter ad content restrictions, launch iing new tools to improve ad transparency and requiring documentation from political ad buyers. we're building artificial intelligence for bad actors. we're working more closely with industry to share information on how to identify and prevent threats so that we can all respond faster and more effectively. and we're expanding our efforts to work more closely with law enforcement. we know bad actors aren't going to stop their efforts. we know we'll all have to keep learning and improving to stay ahead of them. we also know we can't do this alone. that's why i want to thank you for this investigation. we look forward to the conclusions you'll ultimately
share with the american public. i look forward to your questions. >> chairman, members of this committee. twitter understands the importance of the committee's inquiry into russia's interference in the 2016 election and we appreciate the opportunity to appear here today. the events underline this hearing have been deeply concerning to our company and the broader twitter community. we are committed to providing a service that fosters and facilitates free and open democratic debate and that promotes positive change in the world. we are troubled by reports that the power of twitter was misused by a foreign actor for the purpose of influencing the u.s. presidential election and undermining public faith in the democratic process. the abuse of our platform to attempt state-sponsored manipulation of elections is a new challenge for us. and one we are determined to meet. today e we intend to show the committee how serious we are about addressing the new threat
by addressing the work we are doing to understand what happened and to ensure that it does not happen again. at the time of the 2016 election, we acted on instances of automated and malicious activity as we learned more about the scope of the broader problem, we resolved to strengthen our systems going forward. elections continue all the time. our first priority was to do all we could to block and remove malicious activity from interfering with the experience. e we created dedicated teams to enhance the quality of information our users see and to block malicious activity whenever and wherever we find it. those teams continue to work every day to ensure twitter remains a safe, open, transparent and positive platfo platform. we have launched a retrospective review to find russian efforts to influence the 2016 election through automation, coordinated activity and advertising. while that review is still
underway, we have made the decision to review and share what we know today in the interest of transparency and appreciation of the urgency of this matter. we recognize that our findings may be supplemented as we continue to work with the committee staff and other companies, discover more facts and gain a greater understanding of these events. my written testimony details the medology and current findings of our retrospective review in detail. we studied tweets from september 1st to november 15th 20rks 16. during that time we found automated and coordinated activity of interest. we determined that the number of accounts we could link to russia and were tweeting election-related content was comparatively small, about at the time we studied. one-third of 1% of the tweets people saw came from russian-linked automated accounts. we did, however, observe instances where russian-linked
activity was more pronounced and have uncovered more accounts linked to the russian-based internet research agency as a result of thour review. we have also determined that advertising by russia today and seven small accounts was related to the election and violated either the policies in effect at the time or have since been implemented. we have banned all of those users as advertisers and we will donate that revenue to academic research into the use of twitter during elections and for civic engagement. we are making meaningful improvements based on our findings. last week we announced industry-leading changes to our advertising policies to help protect our platform from unwanted content. we are also enhancing our safety systems, sharpening tools for malicious activity and increase eing transparency to promote public understanding of all of these areas. our work will continue for as long as malicious actors seek to abuse our system and will need to evolve to stay ahead of new
tactics. we have heard concerns about russia's actors use of twitter to disrupt the election and about our commitment to addressing that issue. twitter believes that any activity of that kind regardless of magnitude is unacceptable and we agree we must do better to prevent it. we hope that our appearance today and the description of the work we have undertaken demonstrates our commitment to working with you, our industry partners and other stake holders to ensure that the experience of 2016 never happens again. cooperation to combat this challenge is essential. we can defeat this shared threat alone. as with most technology-based threats, the best approach is to combine ideas to increase our collective knowledge. working with the broader community, we will continue to test, to learn, to share and to improve so that our product remains effective and safe. i look forward to answering your
questions. >> thank you. mr. walker, the floor is yours. >> thank you very much, chairman. i'm the senior vice president at google. i oversee our legal, policy, trust and safety and google.org teams. i have worked at the intersection of technology, security and the law for over 25 years. started my career as an assistant u.s. attorney for the department of justice. focused on technology crimes. let me start by conversation with you today by joining the comments acknowledging the victims and families of the awful attack in new york yesterday. as a new york employer, we know how strong and tough new yorkers are and we look forward to doing anything we can to help. turning to the issues before the committee today, google believes that we have a responsibility to prevent the misuse of our platforms and we take that very seriously. google was founded with a mission of organizing the world's information and making it accessible and useful. the abuse of the tools and
platforms we build. goopg l is deeply concerned about attempts to undermine the elections. we are committed to working with congress, law enforcement, others in our industry and the ngo committee to strengthen protections around elections, to ensure the security of users and to help combat disinformation. we recognize the importance of this committee's mandate and e we appreciate the opportunity to share information and talk about solutions. disinformation and propaganda campaigns aren't new and involve many media and publications over the years. for many years we have seen attempts to interfere with our online plat tomorrows. we take these threats seriously. we built systems and put those tools into our consumer products as well. back in 2007, we launched the first version of our safe browsing tool to protect users from malware and other attacks. today safe browsing is used on more than 3 billion devices
worldwide. if we suspect that users are subject to government-sponsored attacks, we warn them about that. and last month we launched our advanced protection program, which helps protect those at greatest risk of attack like journalists, business leaders and politicians. we face motivated and resourceful attackers and continuing evolve iing tools to stay ahead of ever changing threats. our tools don't protect our network security. they also detect and prevent attempts to manipulate our system ises. google news, we use fact check labels to spot fake news. google search, we have updated our quality guidelines and evaluations to help surface more authoritative content from the web. we have updated advertising guidelines as well to prohibit ads on sites that misrepresent themselves. and on youtube, we employ a sophisticated spam and security breach e detection system designed to detect anomalous
behavior. we will continue to adapt to understand and prevent new attempts to misuse our platfo platforms. with respect to the committee's work on the 2016 election, we've look ed across our products to see whether they were using our products to interfere with the u.s. election. we did find deceptive activity associated with suspected government-backed accounts, that activity appears to have relatively limited. any activity is more than we would like to see. we have provide the relevant information to the committee and issued a public summary of the accounts and will continue to cooperate with the investigation. going forward, we will continue to expand our use of cutting edge technology to protect our u users and will continue working with governments to ensure that our platforms aren't used. we will be making advertising more transparent, easier for users to understand and even more secure. in 2018 we'll release a
transparency report showing data about who is buying ads on our plt form and how much money is being spent. we'll pair that transparency report with a database available for public research of election and ad content across our ad's products. we're also going to make it easier for users to understand who bought the ads on our networks. going forward, users will be able to easily find the name of any advertiser running an election ad on search, youtube, or the google display network. we'll continue enhance iing our existing safeguards to ensure we perm only u.s. nationals to buy ads. we already restrict which advertisers can serve ads to audiences based on political leanings. we'll go forward by verifying the identity of anyone that wants to run an ad or use our political interested based tools and confirming that that person permitted to run that ad. we can't do this alone. we'll work to better protect the ecosystem and even as we take
our own steps, we remain open to transparency. manufacture over our commitment to addressing these issues extends beyond our service was. we offered briefings and introduced tools designed to help election websites and political campaigns to protect themselves from unauthorized account access and other digital attacks. we're also increasing our long standing support for the bipartisan democracy project. let me conclude by recognizing the importance of the work of the committee. our users and creators must be able to trust in their security and safety. we share the goal of identifying bad actors who attempted to interfere with our systems and abuse the electoral process. we look forward to continued cooperation with the members of this committee and fellow companies to provide access to tools and help citizens express themselves while avoiding abuses. thank you again for the opportunity to tell you about our ongoing efforts. we look forward to our continuing work with congress on
these issues and we're happy to answer any questions you might have. >> thank you for your testimony. sort of set the stage for what we'll talk about today. as an example, i'd like to highlight one specific case with real world implications involving two different facebook groups. both of which are associated with the russian internet research agency. you'll see the first board that is up. first group is called the heart of texas. this account promoted a pro texas causes and included posts. many would characterize as anti-immigration or anti-muslim. the tag line for this group is referenced in the top left-hand corner of the first chart. texas home left-hand of guns, barbecue and your heart. time to secede.
then 328,000 followers. this account claimed to be pro-islamic themes. the tag line for this group is referenced in the bottom right corner of the first chart. i'm a muslim and i'm proud. if i could have the second board up. the heart of texas group created a public event ob facebook to occur at noon, may 21st, 2016 at the islamic center in houston, texas, to stop, quote, the islamization of texas. the same group then placed an advertisement on facebook to promote their event with over 12,000 people viewed. turning to the second group, it's called the united muslims
of america with over 328,000 followers. this account claimed to push pro-lislamic themes. the tag line is referenced at the bottom. i'm muslim and i'm proud. the united muslims of america created an event on facebook to occur at noon, may 21st, 2016, at the islamic center in houston, texas, to, i quote, save islamic knowledge. same time, same place as the heart of texas event. the group then placed an advertisement targeting people in houston, texas, area to promote the islamic center. more than 2700 people viewed this ad. if i could have the third board. on may 21st, 2016, local news captured the events as they unfolded. reporting on the protest stage by the heart of texas group and resulting counterprotest. the pictures you see on the third board are from the streets in front of the islamic center in houston, texas. what neither side could have
known is that russian trolls were encouraging both sides to bat until the streets and create division between real americans. ironically one person who attended stated, the heart of texas promoted this event, but we didn't see one of them. we now know why. it's hard to attend an event in houston, texas, when you're trolling from a site in russia. establishing these two competing groups, paying for the ads, causing this it disruptive event in houston cost russia about $200. mr. stretch, you commented yesterday that your company's goal is bringing people together. in this case, people were brought together to conflict and facebook enabled that event to happen. i would say facebook has gail failed their goal. from a computer in russia, these operators can create and promote events anywhere in the united states and attempt to tear apart our society.
i'm certain that our adversaries are learning from the russian activities and even watching us today. simply put, you must do better to protect the american people and frankly, all of your users from this kind of manipulation. my time can start now. i have one simple question, yes or no, from each of you. mr. stretch, i'll work to my left. the federal election campaign act prohibits any foreign national from spending funds in connection with any federal, state or local elections in the united states. doesn't this law prohibit your publication of this content? >> prohibit publication of the content we have seen? >> does fec law apply to facebook? >> certainly, fec law, yes. >> it prohibits foreign dollars influencing election. >> it prohibits foreign actors
from using really any medium including facebook, to influence a u.s. election. >> fec law applies to facebook. >> yes, it does. >> it applies to twitter as well. >> yes, sir. >> the prevalence of social media use among military members who spend so much time outside the country deployed away from friends and family seems a likely target for foreign intelligence agency ises who want to collect details on u.s. force movements, deployments and other sensitive insight. do you monitor your platforms for indications that your users in the military are targeted in any way? >> yes, i would say that sort of security work really falls into the traditional cyber security work that we have long been focused on. we had a threat intelligence team for years now focused on tracking foreign actors. it's exactly that sort of threat
that we believe has historically been an area of focus for our adversaries and an area of focus on the defensive side. >> we have been focused on that type of threat for years. we're also focused on education on the other side and helping law enforcement and military personnel understand how to use twitter and its benefits and its risks. >> mr. walker? >> we have been looking at cyber espionage for years. this is all in focus. we're not a social network, we may not have as much visibility whether individual users, but that would be an area of concern. >> these questions are for facebook, mr. stretch. in a blog published in 2017, a facebook chief security officer wrote the company discovered about 3,000 political ads have been paid through for 470 fake accounts and pages and likely operated out of russia. facebook shut down these
accounts on the ground they were inauthentic. had these accounts not violated fib's prohibition against fake accounts, would they have been shut down? >> many of them would have, because many of them violated other pollties related to the type of concept that's permitted on the platform. the authenticity issue is is key. referring to the content you surfaced earlier, it pains us as a company, it pains me personally to see that our platform was abused in this way. people in this country care deeply about issues of public concern. it's one of the strengths of our country that people are so willing to speak freely about them. that foreign actors were able to use our platform to exploit that openness is a deeply painful lesson for us.
one we're focused on learning from going forward. >> does it trouble you it took this committee to look at the authentic nature of the users and the content? >> senator, we are certainly troubled, more than troubled, by the evidence of abuse of our platform during 2016 and we're certainly grateful for the committee's investigation and the attention you're bringing to this issue. we think it's very important. we do believe it's a larger issue than any one company. and we believe that going forward, there are opportunities not just for us to do better, but for us to work together to make sure we're all addressing this threat appropriately. >> what characteristics would indicate that an account or page is likely operated out of russia? >> there are a number of characteristics that can signal
potential location. the most obvious one that is typically the most reliable is location information that's transmitted by the user's browser when they access facebook. they will suggest location, but because of the way the sbirnt is architect, it can also be fake ed. our job is to lock not just for the signals in plain sight, but understand how they can be manipulated and look for patterns of activity that reveal efforts to abuse the plat tomorrow. >> vice president of twitter wastated that twitter is is expanding its resources and building new tool s s to proces bots. what's twitter's process for
identify iing a bot b. >> we have a lot of data behind the things you see on twitter that looks at the activity of an account and remember there are hundreds of millions of accounts. the activity of accounts related to other accounts. our activity looks normal. thousands of times an hour or logs in thousands of times a day it looks suspicious. so is our technology is looking for that anomaly, that normal accounts, but bad actors are getting better at making themselves look more real. >> what percentage of accounts on twitter are actually bots and not real people. >> we do a monthly audit of this. it determined that for years
less than 5% of our accounts are false accounts or spal. >> what happens to accounts suspended on twitter? is there an indefinite status? >> once we us suspend an accoun especially an automated account, they are perm innocently banned from the platform. we do work to link those accounts with new accounts that may pop up. the more we investigate and look into this and build the web of information around the signals we're seeing from these accounts, the better we get at linking those accounts and stopping them before they get on the platform. >> my time is expired, but i'm going to ask you to submit in writing for the record twitter's atssessment of why independent assessments of the number of bots on twitter constantly, consistently are higher than the 5% that you have stated today. >> happy to address that. >> i also want to demonstrate, but also we're get.
ing ready we have had testimony before this committee from a representative of nato that fake accounts on twitter are more in the 12 to 15% account. a vast number of studies ab act of twitter accounts. you're still talking 30 plus million potential accounts. >> this is an example of how people are lured in. first ad, an ad that's pretty beni benign. 217,000 followers. you like that page and he's what happens. you would get a series of benign
bible quotes or other items. this ad appeared in october of 2016. and suddenly this benign site in addition to your bible quotes were getting these other posts, not paid ads, but posts from this organization. this message, the bottom one, would have gone to the 217,000 followers. we have no idea how many times it was liked or shared with other individuals. tools of people lure in and once they lure into a pro texas or pro jesus account and then they are manipulate d by foreign actors. go ahead and start my time. i hear all your words.
but i have a little bit more than frustration. that many of us have been raising this issue since the beginning of issue. and our claims were frankly blown off by the leaderships of your companies and dismissed that there's no possibility. nothing like this happening. nothing to see here. it bothers me if you really committed to trying to work with us to resolve this that it took until this committee continually went at you. it was july and early august when you made your first presentations. your first presentations were less than sufficient. and showed in my mind, a lack of resources, a lack of commitment and a lack of genuine effort. candidly, your companies know more about americans in many ways than the united states government does. the idea that you had no idea of
any of this was happening strains my credibility. so my first question is is this and i want a yes or no answer, not a filibuster. will you commit to continue to work with this committee to provide additional information and additional documents as needed as we continue to explore this challenge and threat on a forward basis. >> yes. >> absolutely. >> absolutely. >> next, one of the things i continue again, and i will commend you here, that from the first friends at facebook, you identified 470 accounts, 3,000 ads and most of the work, at least it appears to e me, has all been derivative of that initial data dump. this is a yes or no question. do you believe that any of your companies have identified the full scope of russian active
measures on your platform? yes or no? >> senator, our investigation continues. so i would have to say no, not with certainty. >> no, we're still working on it. >> we have done a comprehensive investigation, but these are ongoing issues and we continue to investigate. >> let me start with facebook here. you have identified 470 accounts from one troll farm in st. petersburg. there have been plenty of press reports of other troll farms in russia. reports of other activities that were russian-controlled in central europe and eastern europe. in meetings with your leadership as you became more aware of this problem, aggressively promoted the fact that you took down 30,000 accounts around the french elections. now you say not all of those were russian related. have you gone back and cross
checked those accounts you took down in france to see if they were active in the american election. >> senator, the 30,000 accounts that we took down -- >> the accounts related to russian accounts that you took down, your leadership bragged about how proactive you were in the french election process. did you check those accounts to see if any of them were active in the american elections? >>. >> the system that ran to take down those accounts, which were fake accounts of really all type and for any purpose, is now active worldwide. >> just please answer my question. have you reviewed the accounts you took down in france that were russian related to see if they played any role in the american election? >> i apologize, i'm trying to answer the question. >> the answer is yes or no. i want to know have you done --
i have been signaling this to you for some time. we wanted to make sure you would review those accounts. we wanted to make sure the 470 accounts that paid for the 3,000 ads, you said these were all accounts that were paid for in rubles. did you run those accounts to see if they were paid for with dollars or euros or other currencies? >> yes, we are looking and have looked at every possible indication of russian activity in the 2016 election and the investigation continues. that includes any evidence we have identified from those 30,000 accounts and a number of. >> all those accounts have been run to see if if those accounts were active in the united states? >> i will have to come back to you on that, senator. >> sir, we had this hearing scheduled for months. i find your answer very disappointing. on the question of we just
discovered you had 80,000 views in terms of russian views on facebook. we discover in the last 48 hours 120,000 russian-based posts on instagram. have you done any similar analysis on those 120,000 posts? know the 80,000 reached 126 million americans. have you done the same analysis on the 120,000 posts on instagram? >> yes, senator, we have. >> how many americans did those touch? >> the data on instagram is not as complete, but the data we have indicates that beginning in october of 2016, those instagram posts reached an additional 16 million people in addition to the 126 million people that we identify. >> now we're seeing the russian activity is roughly at 150 million americans without knowing how many times they were e reshared. >> the time period prior to
october 16 where the data is less reliable would yield an incremental 4 million. so all told, that gets to 150 million million. >> on the twitter account, there was one activity. this was not something that happened during 2016. i agree we're not here to relitigate 2016. there was a fake tennessee republican account. this account had 154,000 followers. the real tennessee gop party had 13,000 followers. i find very interesting there have been some people who said people should be able to spot these fake accounts. if they are able to spot these b fake accounts, you had the president's communications director e retweeting this account. my question is, why did it take so long to take this down when
the republican party was asking you repeatedly? >> that was an absolute miss. we have refined our policies. >> let me just close with this. my time is about up. >> we have looked on this subject and political informati information, but the same way these farms and fake pages groups, can be used in politics, these same tools can and have been used to assist financial frauds, stock schemes, i think there's a lot of this activity and broad-based digital advertising. we have seen some of this in schemes to sell counterfeit prescription drugs as well as the ability to encourage folks to download malware. i believe this is a real challenge and to get this right we're going to need your ongoing cooperation. thank you. >> thank you for coming today. by now it's probably pretty
obvious to everyone that this committee has spent lots and lots of time on this. both as it relates to the election and these kinds of things not related to the political process in the country, but we have spent a lot to have time. i think they have been able to reach conclusions. they tried to do exactly what was done. >>s there was no success. just like texas, it had absolutely nothing to do with the 2016 presidential election. service it was simply an ak ro moin they were attempting to promote. the chairman talked about the
news reports that alleged that the russians use social media to promote a particular candidate. may even some of those suggest that the change the result of the election. but this whole thing goes a lot deeper than that. one of the things we have discovered and i think you probably are aware of this that you can't look the at those ads and they are promoting one particular candidate. there were ads going both ways, for and against both candidates, but the russians. and i'm going to get back to that in just a second. but the other thing that we have come to the conclusion on and very early, is that the u.s. isn't the only ones that have suffered from this. european, france, germany, just amongst others have suffered tr the exact same thing and that is russian attempted interference with their domestic affairs.
i put a section in the sanctions bill that requires the executive branch to do a study on this. because they were much more overt in europe than they were year here. most of the work they did was covert. and probably because in european countries, there's actually a fair amount of russian sympathy where they can mobilize these people. not so much in the united states. some, but not nearly away there is there. so i want to come at this from a different perspective. the 2016 elections got a lot of the politicians riled up because it went after the political process. but my conclusion is, think most people would agree with me, and indeed senator warner referred to this. this is a lot deeper than just the elections. there are a lot of things that the russians are trying to do. and not just inject themselves
into the electoral process. it seems to me that after you step back and look at this and stay what's going on here, whags the motivation, what are they doing? i look at something from an objective standpoint. what's their objective? what are they trying to accomplish? you walk away shaking your head because americans don't think the same way we think about promoting our country. so the conclusion, i have reached, is that the russians are doing what they have done all along. long before your technology even existed. that is trying to sew discord. and so my question to each of you is have you tried to analyze what the russians were trying to accomplish here not only in the 2016 elections, but in these other kinds of ads with the discord. or your personal views on that.
what they are trying to accomplish. >> senator, it's very difficult for us to determine motive. it's why this committee's work is so important. we have tried to provide you as much information as we can and we hope that with your visibility into other sources of information, you'll be. able to help the american people have a better assessment of what the motive is. we think that will help all of us do better to prevent this sort of activity in the future. >> would you agree that the motive isn't obvious here given the difference in the way they handled these things? >> yes, i would agree with that. >> i would agree with that as well. based on what we have seen, the advertisements from russia today, the types of content that was being. put out by the ira also the aud
mated content looks as if it's merely focused on divisiveness, but we're still investigating this issue and look forward to working with this committee to help put the whole picture together. >> the large majority of the material we saw was on the socially divisive side rather than electoral advocacy. >> that's really been the focus. this was all about the 2016 election. you agree with me that this is much broader than that. and is, as you say, divisive or aimed at discord. would you agree with that? >> yes, that's a problem we're trying to tackle every day. >> yes, i would agree. note that the time period in question and the activity we saw even continued after the election. >> mr. walker? >> that seems reasonable. hard for us to know. >> i appreciate that. as i said, my view of this is this is a whole lot broader than simply the 2016 election. mr. walker, i have a specific question for you. i heard you say that you're
enacting a policy where only a u.s. national can buy an election ad. is that correct? >> that's correct. >> what about other countries? obviously, you operate on places other than the united states. can a national buy an ad for french or german campaign? >> i haven't stated the laws of individual companies, wu we are looking at other elections to make sure whatever we can to mine miez electoral interference. >> so you're going to try to confine people to the own elections in their own countries. is that pretty much it? >> that's going to be a good challenge for you. but good luck. i wish you well in that endeavor. thank you, mr. chairman. >> senator feinstein? >> thanks, mr. chairman. i sat in on the judiciary hearing yesterday.
it was a subcommittee hearing. was able to ask some questions. i want to just make a personal comment because i have been very proud, and i know senator harris is as well, to represent this tech community from california. but i must say i don't think you get it. i think the fact that your general counsels, you defend your company, that what we're talking about is a cataclysmic change. what we're talking about is the beginning of cyber warfare. what we're talking about is a major foreign power with the sophistication and ability to involve themselves in a presidential election and sew conflict and discontent all over this country. we are not going to go away, gentlemen. and this is a very big deal. i went home last night with
profound disappointment. i asked specific questions. i got vague answers. that just won't do. you have a huge problem on your hands. and the united states is going to be the first of the countries to bring it to your attention and others are going to follow, i'm sure, because you bare this responsibility. you have created these platforms. and now they are being misused. and you have to be the ones to do something about it. or we will. and this committee is intelligence. it's different from yesterday. so they are privy to different facts. they are very potent facts. let me go back to a couple of questions that i asked yesterday. yesterday you testified that
twitter only began to remove voter suppression posts that told people they could vote by texting or tweeting after you found out about them from other twitter users. these were illegal tweets waiting for users to alert twitter isn't sufficient. i'll give you another chance. what is twitter doing to proactively identify illegal voter suppression tweets? >> thank you for letting me address that. we are constantly improving not only ob our technology trying to amplify these messages, but also on putting people in in technology on the content and behavior. and trying to make our work flows, our reporting flows more efficient and using artificial intelligence to prioritize things like the illegal voter suppression ads and other things we see on the platform.
and taking those down faster. we are getting better, but this is a problem we are focused on getting better at every day. >> you have to find a way to prevent them from going up. >> we tend to focus on behavior behind the accounts to know before the concept goes up. we have seen great strides in other areas not related to that. we're trying to take that solution to this problem. >> mr. walker, i asked your colleague yesterday why google didn't immediately revoke russia today's preferred status after the intelligence community determined and publically stated that rt was a part of the russian government's efforts to interfere in our election. he told e me that rt only lost its preferred status because of a, quote, drop in viewership, end quote. not because it was part of the kremlin's propaganda machine. this response was deeply troubling and, frankly, did not
answer my question. so here it is again. why didn't google take any action regarding rt after the intelligence community assessment came out in january of 2017. >> senator, let me start by responding to your initial comments to ensure we have taken this issue seriously. the question of cyber espionage is one we have been working on for some years, publicly and privately, working with other companies and working on our own to identify these threats. this is one manifestation of that, but not the only one. with regard to rt, we recognize that concerns that have been expressed by rt and the slanted coverage. thises a question that goes beyond the internet. they have covered the channel on major television stations, advertising appears in newspapers, magazine, airports, it's run in hotels and pretty
much every city in the united states. we have carefully reviewed the content of rt to see that it complies with the policies that we have against hate speech, violence, et cetera. so far, we have not found violations, but we continue to look. the key to this area is transparency. if americans should have access to information for a wide variety, but they should know what they are getting. so we provide information about the government-funded nature of rt. we're looking on ways to expand that to youtube and potentially other platforms. >> i'm really not satisfied with that. that's sort of in the trend of the testimony all along. i think we're in a different day now. we're at the beginning of what could be cyber war. and you all as a policy matter have to really take a look at that and what role you play. i think my time is almost up.
let me try one more. a british report recently concluded that it social immediamedia platforms such as facebook, twitter and youtube failed to remove extremist material posted by banned jihadist and e knee owe nazi groups even when that material was reported. the source is the british parliament's home affairs select committee. last night we saw a horrific attack on innocent people in new york by an individual who may have been radicalized online. we know one person who is a -- 75,000 hits, the major radicalizer in the united states on the internet. i'm working on legislation to require tech companies to report known terrorist taactivity on player platforms to law enforcement and to provide law
enforcement with civil injunction authority. so thank you, mr. chairman. >> thank you. >> thank you for all for being here. i want to as it relates to all of this. it's a whistleblower and critic of the chinese government. his facebook account was blocked and facebook has informed us that he violated terms of service. i think he published personal identifying information about individuals and that violated the terms of service. i understand that argument. my question, what i want to be clear is was there any pressure from the chinese government to block his account? >> no, senator, we reviewed a report on that account and analyzed it through regular channels using our regular procedures. the blocking was not of the account in its entirety, but. i believe was of specific posts that violated our policy.
>> you can testify that you did not come under pressure from the chinese government or any of its representatives or people working for them to block his account or to block whatever it is you blocked. >> i want to make sure i'm being precise and clear. we did receive a report from representatives of the chinese government about the account. we analyze d that report as we would any other and took action solely based on our policies. >> facebook is is not allowed to operate in china. is that correct? >> that's correct. our consumers services are blocked in china. that's correct. >> there have been press reports that facebook may have potentially sdwoped software to suppress posts from appearing in people's news feeds in specific areas. speculation is it's being done for purposes of getting into the chinese market. is that accurate? has facebook developed software to suppress posts from appearing in people's news feeds in
specific geographic areas? >> senator, as you though, we are blocked in china, so any software we have is certainly not ob rative there. we do have many instances where we have content reported to us from foreign governments that is illegal under the laws of those governments. so a great example of this is holocaust denial in germany, for example. our position is if there is content that is visible in a country that the violates local law, we deploy what we call geo blocking so the content will not be visible in that country. >> criticizing a government is illegal in that country, you have the capability to block them from criticizing the government. >> we have the capability to ensure that our service supplies with local law. that's accurate. we take a very nuanceuanced app.
we believe our mission is to enable people to share and connect. >> what if political expression is illegal in the country? >> in the vast majority of cases where we are on notice of locally illegal content, it has nothing to do with political expression. it's things like blasphemy in parts of the world that prohibit blas blasphemy. >> is a foreign influence campaign a violation of the terms of service of any of the three countries here today? if you can prove that someone is doing it on behalf of a foreign
government seeking to interfere in an election, does that violate your terms of service? any of the three companies? >> generally it would violate a number -- we don't have state sponsored manipulation of elections as one of our rules. but the other types of rules like inflammatory ads content would take down most of these posts. so we don't out right ban it. >> i've read that you can buy a bot army from between 45 to $100. if you can prove someone has put together a bot army, would that be a violation of terms of service? >> those would violate our terms of service around the use of automated accounts. we're blocking 450,000 suspicious log-ins a day. we have terms of service
around -- >> is that a violation of terms of service to buy for a foreign influence company or fake ads? >> that campaign violates our terms and policies in a number of ways. and we do not permit automated means for accessing the site. using the bots would be a violation. >> if someone goes on and posts social security number and date of birth of an individual, that's a violation of terms of service, correct? >> for facebook, it is, yes. >> what if someone posts classified information illegally obtained that threatens the lives of individuals or methods or disrupted the ability to disrupt a plot that can endanger the lives of the people. is posting that a violation of your terms of service? it happens sometimes. >> we work with law enforcement all the time on matters like
that and balance free speech rights obviously with those -- obviously an imminent threat we would take very seriously and act on right away. >> it's also illegal to steal and reveal classified information. i'm just curious if that violates terms of service in the fact that it could have real life implications on individuals. do we have any evidence that russian accounts uploaded u.s. voter registration data and used it in conjunction to target specific voters by name? do any of you have any information that registered voter data was used to customize advertising or messaging to individual voters? >> we haven't seen evidence of that so far. >> same is true for facebook. >> the scope of this is not
limited to 2016 or even the presidential race. as an example, i think we've identified being patriotic, lgbt united, heart of texas, all were used to attack my campaign during the primary. on the 3rd of july and the 8th of august one of those, lgbt united attacked again. my point being these operations, they're not limited to 2016 and not limited to the presidential race and they continue to this day. they're much more widespread than one election. it's about our general political climate, that is correct? >> i would certainly agree with that statement, senator. >> with the current fascist leadership of russia undermining our democracy, america must
aggressively confront this espionage and the enemies who sponsor it. the tools are ads, sock puppets and rallies to protests to marches all presented under false pretenses. while the supreme court has ruled that congress may place some limits on strictly political advertising, the other activities i just mentioned are beyond the reach of government and government regulation in a free society. to fight back against this espionage, americans have to rely on our marketplace of ideas and the institutions that support it. gentlemen, today you three represent those institutions. you've discussed your response
to these attacks, but it is self-evident in relation to the power your platforms now have in the past election you failed. and this is especially troubling, because the same federal law that allowed your companies to grow and thrive, the section 230 law gives you absolute legal protection to take action against those who abused your platforms to damage our democracy. thegorithms that power your companies including fake accounts and news stories and identify the source of money purchasing your ads. now, i'm of the view ads are a small part of a much bigger problem. fake users posting stories on facebook, videos on youtube, links on twitter can be used by
foreign and domestic enemies to undermine our society. you need to stop paying lip service to shutting down bad actors using these accounts. you've got the power and congress has given you the legal protection to actually act and deal with this. so i want to start with a couple of quick yes and no questions. mr. walker, are you satisfied with your platform's response to foreign interference in the 2016 election, yes or no? >> we are constantly doing better. >> is the answer no? >> we could have done more. >> i'll take that as a no. >> no, we need to do more. >> mr. stretch? >> same is true. >> do you all have and we'll start with you mr. walker, the
technical ability to respond to future campaigns? >> yes. the safe harbors and -- >> yes or no? >> yes. >> stretch? >> yes. and i would add, though, that i do believe we need information sharing among industry as well as working with the government to enable us to do this effectively. >> gentlemen, specifically now describe the changes you're going to pursue that respond to not just the ads, but the sock puppets, the hoaxes. we'd tlilike to walk out of her knowing the changes you're going to support. >> the transparency report that we talked about for ads, an archive of content that's available. i cons that make information on
the site available to users as to who sponsors an ad and enhanced verification techniques. when it comes to fake news, we're improving our algorithms and signals we use. we're using fact check labels to improve users' ability to evaluate fake news. and we're looking at our ad policies. >> coming out of the 2016 election and early this year, our ceo asked our entire engineering product and design teams which make up a large majority of the company to tackle the problem of safety abuse and misinformation on our platform and to drop everything else that we're doing and to figure this out. we formed an information quality team -- >> those are three sentences. what are the changes? >> we formed an information quality team focused on looking at behavior and content and
seeing how we could stop bad actors from using automated activity. we just announced new transparency rules around not just political ads but all advertisements to educate not just american citizens but our world wide users. we also continue to collaborate with law enforcement. >> i know very few specifics from that answer. mr. stretch? >> let me try four things. first today there are 10,000 people working at facebook on safety and security security product and community operations teams. by2018, there will be more than 20,000. second, we announced last week a series of ad transparency steps drawing on the ideas and the honest ads act that senator warner talked about earlier that will bring much greater visibility to advertising generally and particularly to political advertising.
third, we are tightening our ad policies to limit divisiveness and to limit violence in the use of our ad tools. fourth, we're standing up an organization to enable better industry sharing of threat information and also to help us work better with law enforcement so that we can share information in order to address this threat. >> my last question is, it's not clear that you all or the public understand the degree of this sophisticated and manipulative intelligence operation. the russians created facebook pages, posted youtube videos, all trying to appear to specific audiences. some of the content wasn't fake. it was intended to gather an audience and gain trust. it told people that they were already receptive to that after gaining that trust, you could execute the espionage, for example, by gathering liberals and then discouraging them from
voters. mr. stretch, i'd like you to confirm that this technique was used in the election. >> senator, we've provided all the information we can about the content that we've identified on the system. i think to make the sort of assessment you're describing really requires this committee's work to look at all of the online and offline activity that would be necessary to effectuate a campaign like that. >> my time has expired. we have specific cases that that was used. i would like in writing within a week what you're doing about it. thank you, mr. chairman. >> senator collins. >> thank you, mr. chairman. it is very clear that russian activities on your social media platforms go far beyond the paid political ads that appeared last year. the primary purpose of russian
active measures is to exploit and aggravate the divisions in american society and to undermine public confidence in our democratic institutions. and those efforts have not stopped. they continue to this very day. as senator rich has pointed out, no area of the country is immune. let me give you an example. we've passed it out to you. by describing three unpaid posts from facebook pages created by the russians that referred to the governor of maine. there are two negative posts related to the governor. on one russian facebook page called williams and calvin that appeared in the august of 2016. there's a video of comments made by maine's governor from that
same month. and the post in part says the following. he called up white people to kill blacks. we can clearly see what kind of people serve in american government. white racist supremacy, that's for sure. the only way to avoid mass killings of black people is to fire lapage and all who have the same racist believes from american government. there was a second post on the same website about ten days later. let me read part of that. it is not a secret that america is the country of white supremacy and people like lapage must be replaced from their positions in the government. america doesn't need racist politicians. black people are tired of white
supremacy. then this year, this year, in august of 2017, maine's governor was the subject of a positive post on a different russian backed facebook page called being patriotic. in this case, the post defended comments that the government made at the time about confederate monuments. the post ends with its own incendiary conclusion. it says when even the governor is not safe from leftist haters, then what can we say about ordinary citizens. liberals are now acting like terrorists. they try to intimidate everyone who disagrees with them. hope our police are take appropriate measures against
these cowards. let me point out something. our governor was not up for reelection last year. he is term limited. he cannot run for reelection as governor. and yet these comments were made both last year and just a few months ago. and the posts are just three among 80,000 that reveal the russian play book of playing both sides off against each other and of sewi isowing disco division with inflammatory red ric -- rhetoric. the russians continue to push this kind of divisive rhetoric to this very day. so my question to you is what
are you as american companies doing to effectively counter unpaid content posted by the russians that is clearly designed to specifically polarize and anger the american people. and i would argue that you have a special obligation here given your reach in american society and the fact that you are patriotic american companies. mr. stretch. >> senator, we agree that we have a special responsibility here. we value the trust that users place in our services. and when they show up to connect with friends and family and to discuss issues, they need to
know that the discourse they see is authentic. what is so painful about this type of content is that it exploits truly and passionately held views and then inflames them to create more discord and more distrust. to prevent this, we are investing much more heavily in authenticity. we believe that one of the cornerstones of facebook is that users are known by their real names. and so that creates a level of authenticity in the discourse that users can trust when they come to the platform. this sort of content erodes that trust and it's contrary to everything we stand for as a company. as americans, it's particularly painful because it is so expl t exploitive of the investment in our society.
the commitment we're making is to ensure our authenticity politician is more effectively policed and monitored to prevent exactly this sort of behavior. >> that is twitter doing? >> the one we see the greatest strides in is on the amplification side in the use of automated accounts. these bad actors need an audience for their voice. and generally they don't have a followership so they're trying to use activity on the platform to amplify their voices. so we're looking behind the message and the content at the behavior of doing that and have been doubling our effectiveness year after year and taking down millions of accounts every single week because they're not actually humans. >> this just happened in august of this year. this isn't something old.
>> right. we continue to try to stay ahead of their activities. we're also looking at things like coordinated human activity where real people are coming together like the ira and putting out divisive content like this. we're able to link those accounts and take action on not just what they're saying, but what's behind it. we've had great strides on the terrorism front. >> mr. walker? >> we're also very concerned about this kind of deceptive and divisive content. we have removed these going forward. we have engaged a number of things to avoid the problem of fake news, changes to our algorithms, labeling of fake news where we can find it working with third parties, et cetera. >> thank you. mr. stretch, i want to start
with you. last month president trump called russian purchased facebook ads a hoax. i've looked at those russian sponsored facebook ads. i certainly hope you've had a chance to review them. are they, in fact, a hoax? >> all the information we've provided to the committee did run on facebook. so -- >> it's a yes or no answer. i know you're a lawyer. it's hard, but -- >> no. the existence of those ads were on facebook and it was not a hoax. >> in the interest of just clearing this up and giving the american people some transparency into this so they can see the nature of what typically gets used to divide the american populace, why not simply release those russian financed facebook ads to the public, redact the pictures but release the content so people
can understand how this works? >> senator, we believe this committee is really best placed to determine what information to release. we stand ready to assist in that effort. we agree that the more people can see the type of content that ran and the divisions that were sought to be exploited, the better. >> i think we have a disagreement on this kmicommitt whether or not to release those. i want to move onto russia's rbc magazine, which recently revealed that st. petersbeurg's troll factory employed hundreds of trolls and spent about $2.3 million in 2016 to meddle in u.s. politics, actually contacted u.s. activists
directly and offered them thousands of dollars to organize protests. your platforms are all global. they're not just u.s. platforms and there are substantial open sourced reporting right now suggesting that similar divisive activity may be occurring, for example in the catalonian region of spain right now. what are you doing right now to make sure that your platforms aren't being used in similarly divisive ways across the globe to sow discord in western democracies? are you familiar with what you're doing there? >> senator, we are focused on preventing this form of abuse globally. so when we say we have an obligation to protect the
platform from being used for abuse, that's a global obligation. so we are focused on elections as they appear on the calendar including the catalonian election that occurred recently as well as the other elections that are on the calendar going forward. we're focused on ensuring that all actors on the platform comply with local law and we are focused on making sure that any foreign threat actors that are seeking to undermine democracy anywhere are removed from the platform. >> have each of you had to take corrective action against actors in that debate who were not who they purported to be? >> senator, the key i'd say progress we've made -- >> that's a yes or no once again. >> i believe so but i'll need to follow up with your staff. >> mr. walker?
>> we're constantly removing fraudulent accounts from our services. >> you can get back to us. given the discussion we've had about automated twitter accounts and bots and the range is obviously very wide but we know that's a problem. and you made an assertion earlier that i want to come back to and just make sure it's accurate. do you require at twitter by service agreement that profiles are linked to real names, real people or so other way to make sure those go back to real human beings from social security numbers to other unique identifiers? >> we do not. we require some information at sign-up, but we don't require you to verify your identity. we have services that verify identities on the platform. >> why on earth not? >> because we see the power of twitter being used by folks like our political dissadents,
embedded journalists in difficult country who is use the ability to not have to identify themselves by name to speak their truths to power. >> so the reason is for social dissidence and people in third world countries where there is a hostile government regime. it is not your business model. you're not reliant on those automated accounts to generate revenue? >> there's some good automation on the platform. but we do not rely on the bad malicious automation that we're talking about here. >> if i were running political campaign today and i were to advertise on local television, on cable television, in print or on the radio or even through the mail, i would have to have a paid for by disclaimer on those ads. now mr. walker i believe has already addressed this issue.
but is there any policy reason that online social media ads given how effective and influential they have clearly become shouldn't meet that same level of transparency? >> we agree with the transparency efforts and last week announced that we're creating a transparency center not just for political ads but a transparency center for all ads so that you can see not just the ad that you've seen and why it's been targeted to you but all of the other ads created from that same advertiser. on te electi you'll be able to see what the targeting criteria are. >> mr. stretch. >> is same is true for facebook. we are working both of political ad transparency, enabling more visibility into campaign ads by third parties and also enabling
campaigns to meet their disclosure obligations in their online communications. >> so in response to mr. hi hienrick's question, there was a lot of information that you could get based on that policy. would you get that by going to another spot? surely that's not all right there on the ad. >> obviously we're a character contained platform. we will be identifying very clearly whether or not something is a political ad so you can see it right away. and then depending on the uh yoe on a mobile phone you'll have to hover over to see a full transparency center. >> would you be able to on the ad itself to put enough of disclosure there so that it's clear when you're looking at the ad who paid for it and how to
find more information about who paid for it. >> we're still working on the technical details. >> mr. walker, are you trying to do anything similar to that? >> our idea is to have an icon that a user can click on so it's immediately available to them when they see the ad. >> would any information be on the ad except the icon? >> depends on the format of the ad. the sec has struggled with figuring out appropriate disclo disclosure requirements. >> the same is true for facebook. >> when you're talking about russians, are you referring to the russian government or any russian citizen or people who paid in rubles?
you mentioned the ira, which i assume not either your individual retirement account or the irish republican army. so how do you know they're russians? >> that's a great question. we are looking for signals. not everyone identifies themselves as a russian, especially these malicious actors. do they have a russian phone number, do they have a russian e-mail address. are they coming in from a russian i.p. have they ever logged in at any time from russia. there are some technical challenges with that. the trail sometimes goes cold at data centers where information is being processed. >> looking back at what the russians did during the election, when you're saying the russians paid for these ads,
these are ads paid for by the russians because you've now gone back and checked groups like the research agency, the internet research agency and you now know that's a russian group. >> we also have banking information because folks are paying for these ads. we didn't link the ira accounts to advertising in the election but we found nine advertisers based on the signals that i talked about. largely it was russia today who we've since removed from advertise advertisers. >> i've always had some sense that in advertising you pretty much got what you paid for. mr. stretch, how much money did the russians spend on ads that we now look back as over
disruptive or politically intended? >> approximately $100,000. >> how much of that did they pay befo before the election? i've seen the number 44,000. is that right? >> the ad impressions ran 46% before the election. the remainder after the election. >> 46%? well, if i had a consultant that was trying to impact an election and spent only 46% of the money before election day, i'd be pretty upset about that. so they spent $46,000. how much did the clinton and trump campaigns spend on facebook? i assume before the election. >> combined, prooxly --
approximately $81 million. >> before the election? >> yes. >> so 81 million. i'm not a great mathematician. but 46,000, 81 million, would that be like five one thousandths of one percent. >> it's a small number by comparison. >> the fact that we're talking about it today, certainly seems like they got their money's worth after the election, whether they got it before or not. we're still talking about five one-thousands of 1% on the facebook money spent. was that just clinton and trump or all the candidates put together? >> no. those were the clinton and trump campaigns combined. >> if you're -- i know ferguson,
missouri, and baltimore were a couple of big targets in a lot of these ad campaigns. is there a way you could do this to where you principally target viewers in the st. louis area for ferguson or in maryland for baltimore? >> it's important to distinguish between our ad tools and the organic tools. our ad tools do permit geographic targeting of content. approximately 25% of the ads that we've identified were geographically targeted to a region smaller than the united states. most of them were targeted on a state basis. organic content, unpaid posts, if you will, are not geographically targeted. >> and some of those targeted dollars were spent in states where the election turned out not to be close at all, is that right? >> that is correct. >> the other questions that we'll get to maybe in writing
lat later, but on the free media, i think we have to be very thoughtful here about who decides what's voter suppression and what's not, who decides what level of speech is acceptable and what's not. it's an unbelievable obligation that the government has never been very good at. that will be an on going discussion of whether that's possible or not and the questions and problems that arise when somebody does begin to decide what's acceptable to talk about and what's not and what discourages voters and what doesn't. i would think the general election process these days would discourage voters from participating. maybe that would just mean none of it could be discussed but we will see how that goes.
>> i'm disappointed that you're here and not your ceos, because we're talking about policy and the policies of the companies. it's fine to send general counsel, but i think if you could take a message fwrback fr this committee, we would appreciate seeing the top people who are actually making the decision. i want to gibegin by two quotes. the first one says this, nothing was more to be desired than that every practical obstacle should be opposed to, intrigue and corruption. chiefly, from the desire of foreign powers to gain an improper ascendant in our
counsels. that's alexander hamilton. he saw this coming. the other is a more recent quote from a fellow who is a former gru officer. he said a new type of war has emerged in which armed warfare has given up to another kind of warfare, information warfare. and that is exactly what we're talking about here today. and i think it's -- i appreciate the chair and the vice chair giving us the context of what they're doing. their visual demonstrations were very vivid. the warfare is the division of our society. it's not only us. it's the entire west. we know that the russians were involved in the french election. we know they were involved in the german elections. we now know they're involved in
the separation of spain. we now know they've set up shop in scotland which is talking about an independence vote from great britain. this is a sophisticated world wide strategy that, applied here in 2016. there is one other piece i'd like to add to what the chair and the vice chair did and that is that it's still happening. this is a service of the german marshall fund called interestingly hamilton 68 that follows hashtags on a daily basis. i just picked a day in september to show these are the hashtags that are being propagated -- they weren't created by russians but these are 600 russian websites that are using -- these are the hashtags that they did on these particular days. syria is up there. then we have the nfl and then we
have boycott the nfl, then we have stand for our anthem, we have make america great again, russia take a knee. in other words they were tweeting on both sides of the nfl dispute in order to exacerbate the divisions. one witness to this committee had said that their strategy is to take a crack in our society and turn it into a chasm. that's exactly what we've seen, we saw in 2016. and my point here is, it hasn't stopped and it won't stop. we have to figure out what to do about it. it seems to me there are three possibilities, one of which you can make a significant contribution to. other two frankly are up to us. the first is the technical defense, the kind of thing you've already been talking about today, checking identities.
the second is we as a society have to understand when we're being conned. i spent some time a year ago in eastern europe before our election. and the eastern european politicians, all they wanted to talk about was russian meddling in their elections. i said how do you defend yourself? they said all of our people now get it that this is what the russians are doing. when they see one of these postings, they say, oh, it's just the russians again. we have to develop that level of sophistication so we know when we're being mislead. the mentality is that when we see a tabloid that says a movie star had a two-headed baby, we say, oh, that's just a tabloid. we need to apply those same
sensitivities to these kinds of fake news and purposeful distortions. the third thing is that this country has to have some kind of cyber warfare deterrent capacity. right now there's no price to be paid for meddling in our democracy. and our adversaries have to understand that if they're going to undertake a campaign like this, will will be a price to be paid. there will be results. and if they do x, we are going to do y to them. right now that doesn't exist. and all of what the russians did last year has basically been a free pass. i think that's a very difficult problem. now let me ask the technical question. could you for example require a dateline on a posting that said where it comes from just like a news story says moscow, september 23rd? is there some way to identify the source of information as it comes across your news feeds?
>> senator, it's a great question. we do permit users to identify the geographic location of the post. we don't require it. there are oftentimes privacy considerations that would prevent a user from -- >> you could require it by country, couldn't you? >> there are many uses of our services, senator, where requiring people to designate their physical location could be problematic. i would make two other points. one is because of the way that internet is architected, your geographic location can be disguised. that's something we need to work on in order to make sure we're not being fooled, because i think your larger point is an excellent one, the geography of the location of the user paired with the content they're serving
is part of the information and we need to do a better job tuning our systems to be more sensitive to that. >> mr. walker you said something similar, they the people should know what they're getting. when we get information in the newspaper, we see the author and the name of the dateline. we're in a new information world here and we need to think about how to apply some of the prin principles that help us asesds tha -- assess that information. i hope you continue to develop policies to help your customer to analyze the data. the problem now is we're just taking what it comes as it comes. i have a quote on my kitchen wall that my wife found. it said the great problem with
quotes on the internet is determining whether they're authentic, abraham lincoln. >> thank you for being here. i hope you hear it loud and clear from this committee there are lots of questions. this is not an opposition to free speech, though. this is actually a battle to try to protect free speech. we want to have good american dialogue. the fear is that your platforms are being abused by foreign actors who want to abuse free speech. if two americans have a disagreement, let's walk it through as two americans. if an outsider comes into it, we do have a problem with that. we're grateful that you're here and we look forward to cooperation tailgate to figure out how to actually resolve some of these extremely complicated issues. i do want to push about the type of ads and content. you have mentioned from your testimony that in twitter, of
the 131,000 tweeted posted during the time -- i assume that means that september to november time period that you were tracking -- only 9% of those tweets from those russia targeted accounts were actually election related. i think that's being lost in the conversation that only 9% of the tweets were election related. my question is what are you seeing from russian related accounts that you're tracking now and trying to be able to pull down or identify. what are the soeshcial issues bg discussed right now trfrom thos sites. one has been mentioned. that's the nfl. either boycott the nfl or take a knee on that. has that been actively pursued on your sites?
>> so one example we saw following the election was an effort by the accounts we've identified to inflame some of the post election demonstrations we saw. so some of the accounts turned to questioning the electoral college as an example. >> so we'll walk through multiple examples here. some have reported that the site was used to organize protection events in certain cities, sending out messages saying we're all going to protest at this spot but obviously it was created by a russian group. what else, nfl, yes or no on that? >> we've seen that activity. >> what other issues? >> we've seen more limited use of our services, but among that i would say police shootings and racial issues.
>> what else? >> certainly immigration has remained a topic throughout. >> any other issues? this is part of the reason that multiple members -- we really do want these ads to get out in the public space. we think there's great value for all of your platforms to say this is the type of content that foreign actors are trying to put out that is actually divisive content. we think there's a great benefit for you to be able to say when you're aware of things. please note this is the type of issue that's coming out and this is what it looks like. so people can say that's the type of thing i have seen before or they can say i've actually liked on that before and didn't have any idea that was russian related. when were you aware of russian activities on your platform during the election time?
during or before the election? when were you aware that entities within the russian government related or policy related were involved in election issues on your platform, before or after the election? and all three of you can say. before or after? >> we were aware of russian state actors active on the platform prior to and through the election, separate from the internet research agency. we communicated with law enforcement about our concerns at the time. these actors were engaged in more traditional cyber threat activity, focusing on account compromise as well as trying to direct attention to stolen information that was hosted on other sites. >> how far back? early 2016, 2015? >> we had seen activity as early
as 2015. >> twitter also saw activity from the i.r.a. on our platform and took large scale action to take down accounts in 2015. we generally are aware of intelligence community reports. so became aware of the activity in the report that came out in january of this year. and then obviously through the retrospective work have uncovered what we think is the full extent. we're continuing to look and research that issue. >> mr. walker? >> we've been looking at cyber espionage issues for many years. it was only after january that we took a deeper dive. >> one of the things we're trying to address is getting additional information. getting the type of posts that have been done would be very helpful to the american people to be able to see. the statistics that you've given
us are the number of accounts. brand new it's not giving us the breadth and depth of the accounts. 1% of those accounts were russian related. but what we're not getting -- are they in the middle stirring it and advancing it or were they beginning it? that would be helpful to us as well. just giving us the percentage of users that shared it doesn't help us. knowing when in that process and what happened that is what helps us. y'all have done a lot of work on terrorism, on islamic extremism, on the advance of isis, child pornography, human trafficking, on the sales of illegal drugs. we're asking for help on this area as well. this is something that i would
hope that we do not have to engage legislatively. this is something you have been actively engaged in. i do hope in the days ahead we continue to protect a platform for free speech to allow individuals to speak their opinions whether it's agreed on or disagreed by americans. >> the russian state news organization the federal intelligence officials call the kremlin's principal population outlet. they post derogatory information about the united states. it's information warfare directed against the united states according to the unclassified intelligence report released on the 2016 elections. the kremlin staff's rt coverage recruiting people who can convey russian strategic messaging. the kremlin spends $190 million
a year on the distribution and dissemination of rt programming. do any of your organizations allow r tort to purchase space? >> we've offboarded, which means we've beened russian accounts from advertising on the twitter platform. >> effective as of when? >> as of a week or so ago. >> we have reviewed the rt accounts. we've not found violations of our accounts but we are focused on making sure there is enhanced transparency. >> the same is true for facebook. >> it says here that rt uses google, youtube, facebook and twitter as the main distributors of their content. so you all have been the main
go-to the get their propaganda out to the united states to do harm to us. >> of course rt channels are included on major cable television networks, satellite hotel networks. it is a problem that goes beyond the web. >> their anti-virus software, do any of you all have it on your personal computers? >> no. >> does your company? >> i don't believe so. >> could you check and find out? >> we will follow up. >> not aware, but we'll follow up with your staff. >> likewise. >> do you have any doubts about russian involvement in our 2016 election?
any thoughts than? do you have any doubts about that? >> we have no doubt that there are attempted efforts at interference. it's something we're focused on addressing going forward. in terms of whether it had an outcome on the election, that's something we're not in a position to judge. >> are you or are you ceos concerned about the threats to the u.s.? are you concerned about that? do you talk about the threats to the united states of america where you're domiciled? or is it basically just a business model that uh you're worried about? >> we're da ee're deeply concer. we pointed our entire engineering and design team on tackling the issues. >> i join that.
it's absolutely a very serious issue. the north star of google is to provide accurate, comprehensive, relevant information to people. we don't always get that right but we have tens of thousands of engine engineers who are addressing the problem. we take this very seriously. >> you all would agree then i guess with the legislation that's been introduced by different members of this committee and other committees that you all should be regulated and overseen the same as we do other news medias? anybody have any idea on that comment? are you all going to fight back on that lobby against it or are you going to basically support the legislation that's going to be needed to make sure the american people are getting the facts and not fake news? >> we stand ready to work with the committee on legislation. >> you've seen the legislation
we put out there. you all three are lawyers. i'm sure you're watching it very closely. do you agree with the exception? would you support a change with a law that treats all advertisers the same as print and broadcast ads? >> very supportive of the direction of the honest ads act. >> you all will be speaking in support of these pieces of legislation? >> we have some fine tuning. we put out our own transparency center that very much aligns with the information that the honest ads act was asking us to provide to you. >> according to a study, russian trolls are now targeting american military personnel and veterans on twitter and facebook. one fake facebook page was highlighted in a stars and stripes article from october 18th. the page was called vietnam vets of america.
it had ak tractattract add foll network of nearly 200,000. a week after the original article, it was reported on october 18th facebook reportedly took down the site, ostensibly for violating intellectual property. is the story accurate? and are you seeing our military and veterans being targeted and how come it took so long to take it down? >> we're intensely proud of the use of facebook by our military. >> we're taking about the people targeting them. >> yes. and we are very focused on making sure that it remains an authentic experience for them. we received many reports of inauthentic behavior.
we're trying to act on them quickly and improve our tools to detect it. i'm not familiar with the particular page you describe but i will certainly check into it. >> you can see all of us are very much -- this is not a democrat or republican issue. this is an american issue that we're concerned about, the security of our nation. we're getting hit from every way you possibly can imagine. and you all are the one of the largest distributors of news. there can be no doubt that it has to be authentic and true. you cannot allow what's going on against the united states of america. you are on the front lines with us and we're doing everything we can to support our military, our veterans, all the people who put their lives on the line. what you're doing by allowing this fake stuff to come across, this misleading, this damaging information is really threatening the security and
safety and sovereignty of our nation. i would hope that your ceos would be here. they need to answer for this. it can't be a business model. it's got to be a security issue. thank you. >> senator cotton. >> thank you, gentlemen, for your appearance this morning. mr. edgett, i want to discuss twitter's history of cooperation with our intelligence community. last year, in an open hearing before this committee, i asked then cia director john brennan about twitter's decision to prohipr prohibit a subsidiary from working with our intelligence committee. he stated he was disappointed in twitter's decision, but at the same time, we learned twitter was refusing to work with the cia and the rest of the intelligence community, we also learned that twitter was pitching russia today and sputnik propaganda arms of the kremlin to sell advertisements for profit. so in essence, last year, russia was beginning its covert influence campaign against the united states, and twitter was
on the side of russia as opposed to the national security interests of the united states. how can your company justify this pattern of behavior to its fellow citizens? >> we work frequently and hard with law enforcement all the time. we do have global policies that prohibit the use of our data hoses or publicly available data around tweets for purposes of surveillance. but we allow law enforcement to use data miner and twitter products around news alerts, first response technology to see what's going on in an area, if the 911 call is made and an emergency responder is going somewhere, but we do not allow surveillance based on twitter data. >> did twitter cut off the cia and the intelligence community from data mining last year? >> we asked our policy of surveillance be applied consistently, and that data m e miner has been enforcing that.
>> when we approach russia today last year, to talk about our advertising products and to sell them our advertising services, they were approached as a regular media organization, like bbc or npr. >> do you consider rt to be a regular media organization? >> obviously not now. coming out of the dni report earlier this year and the retrospective work we have done most recently, we don't. that's why we have banned russia today from advertising on the twitter platform. >> so there's a difference, though, between the advertising question, which was improvident and the use of data miner. according to a "wall street journ journ journal" report to which brennan was responding, twitter ceo objected because he objected to the, quote, optics of continuing to help the u.s. intelligence agency. that report also said customers still getting data miner includes rt. is that an accurate report?
john brennan had no reason to doubt its accuracy. >> i don't have the information but we'll follow up on russia today's use of data miner products. i believe mr. dorsey wanted to make sure that our policies were being applied consistently around surveillance. >> do you see an equivalency between the central intelligence agency and the russian intelligence services? >> we're not offering our service for surveillance to any government. >> so you will apply the same policy to our intelligence community that you apply to an adversary's intelligence service. >> as a global company, we have to apply our policies consistently. >> which reminds me of the line from the cold war that of one who did not see a distinguishment between the cia and the kgb on the other hand because the kgb officer pushed an old lady in front of an
oncoming bus, but the cia officer pushed the old lady out from the path of the oncoming bus, because they both go around pushing old ladies. i hope twitter will reconsider its policies when it's dealing with friendly intelligence services in countries like the united states and the uk as opposed to adversarial countries like russia and china. do you -- would twitter entertain the possibility of once again allowing the intelligence community to use data miner? >> we do today, for purposes of news alerts and first response technology, getting information on certain areas. dwee not allow anyone -- our policy is not to allow anyone for the purpose of user privacy, to use our technology to run surveillance. >> okay, let's move on to another hostile intelligence service. other than vladimir putin and russia, i can't think of anyone who is more involved in the
efforts to influence our election last year than julian assange and wikileaks. the current director of the cia, mike pompeo, as well as this committee, has labeled wikileaks a nonstate hostile intelligence service who aids hostile foreign powers like the kremlin. yet, to my knowledge, twitter still allows him to operate uninhibited. is that accurate? >> we have terms of service and rules that apply to all users. we'll apply those consistently and without bias. we take action on accounts like wikileaks. >> is it biased to side with america over our adversaries? >> we're trying to be unbiased around the world. we obviously are an american company and care deeply about the issues that we're talking about today. but as it relates to wikileaks or other accounts like it, we make sure they're in compliance with our policies and have and will continue to -- >> so you'll be unbiased towards wikileaks and julian assange, but you'll take down videos of
people like marsha blackburn, a senator running. >> marsha blackburn's video was never removed from the twitter platform. she ran that tweet and video as an advertisement, and we have different standards for our advertisements than we do for the organic tweets and content on our platform because we're serving ads to users who haven't asked to follow representative blackburn or others. we want to make sure that's a positive experience, and so our policies have a different standard. and in that case, we had users reporting that it was inflammatory and upsetting, and it was initially taken down. we're making these tough calls all the time. in that case, we reversed the decision and allowed the adver tizement to continue to run. we never took down representative blackburn's tweet or never allowed those who were following her to engage in dialogue with her. >> i know you're the acting general counsel, and if you're the general counsel, these are made at the ceo and board of directors level, but i have to say, most american citizens
would expect american companies to put the interests of our country, above, not on par with our adversaries. kuncountries like russia or chi or nonstate actors like wikileaks or individuals like julian assange. as many other members of the committee have expressed, i think your companies have accomplish eed amazing things f our country and citizens. i also support the channels you have created for free speech, especially for some oppressed or persecuted people around the world. but this kind of attitude, i would submit, is not acceptable to the large majority of americans, and it's going to be part of what would lead to unwise or imprudent regulation, not sensible and smart regulation. my time has expired. >> thank you. california's home to many of the world's most successful technology companies and we're proud of that. and we also know that with that
great success comes great responsibility. your companies, therefore, have a great responsibility to the american public. and as you know, you are the modern town square and the modern post master. you are the phone company and the yellow pages. you are the newspaper and the radio broadcaster and the television station. and you are the emergency alert system. your decisions fundamentally inform public discourse. so our nation's enemies have used your platform in a way that has been designed to create and disseminate and advertise hateful rhetoric with the intent and the effect of disrupting our democracy, and that, of course, is why we're all here. i have several questions, but i would like to start with what i think is frankly an elephant in the room. i'm holding up the s.e.c. form 10-q that has been filled out by each of your companies, and your response to this is pretty much
the same, but i have in front of me that from facebook, twitter, and alphabet has the same information. there's a section here which requests what are the risks related to the business? and it reads, if we fail to retain existing users or add new users, that will be a problem for us. it goes on to say, we generate substantially all of our revenue from advertising. the loss of marketers or reduction in spending by marketers could seriously harm our business. it goes on to say, our advertising revenue could also be adversely affected by a number of other factors including adverse legal developments relating to advertising, including legislative and regulatory developments and developments in litigation. so my question to you is, i would like to talk specifically about american ads. not the russian ads. american ads. that run on your platform. there are legitimate ads that
appeared alongside of the russian placement and propaganda pages on facebook, on twitter, and even on youtube. so can you please tell me that as it relates to those advertisements on facebook, on twitter, or in youtube, how you are addressing that, and in particular, how much money did you make off the legitimate ads that ran alongside the russia propaganda? we can start with google please. and that would be the advertisements that ran before your videos on youtube. >> sure, the total amount of advertising we discovered across our platforms was $4,700 from the russian sources. >> that's not my question. legitimate advertising. how much money did you make from legitimate advertising that ran alongside the russian propaganda?
>> i don't have it in front me. we would be happy to follow up. >> what about for twitter? >> i don't have the data, but i'll follow up. >> have you not looked into that? >> i believe are you asking how much advertising revenue we made for the period totally? >> i'm asking how much advertising revenue did you receive from legitimate advertisers that advertised alongside or in connection with russian propaganda? >> we haven't done that analysis, but we'll follow up and work on that. >> okay. what about facebook? >> the same is true for facebook, senator. >> you have not done that calculation? >> we have not done that analysis. >> i find that difficult to understand because it would seem to me we would figure out how much you profited from russian propaganda on your platforms. so please do follow up with isas soon as possible on that. and also, it is critically important that this committee have access to all of the information it needs to understand the russia propaganda
in the 2016 election, so will you each commit to retaining records as you're required to do from the minimum standards of media ratings research, but do that and extend it beyond the 11 months they require and extend it through the completion of our investigation into russia's interference in the 2016 election? will you commit to keeping those records during the duration of our investigation? >> we will keep all relevant records to this investigation and provide them to the committee, yes. >> same goes for twitter. >> thank you. >> and for all three of you, can you please name the senior executive who is responsible in your operation for countering state sponsored information operations? and if you do not have one, please indicate that as well. >> it's a challenging question because we have a number of people across different teams, including our cyber espionage teams as well as our trust and safety teams. i would say our chief security officer is one such person. another person would be the head of our trust and safety team, then we also have separate teams
at youtube. >> i take it you have not -- you have not designated an individual as part of your executive team who is responsible, specifically, for state influenced operations. >> i will take responsibility for that, senator. >> i appreciate that. >> there's two people filling that role at twitter. the first is our general counsel. i'm currently your acting general counsel, so currently it's me. also the head of our twitter product, the twitter product that we all use, has taken responsibility for safety abuse and information quality on the platform, so i feel like that's directly related to your question. >> okay, but i would like you each to appreciate and everyone to appreciate that this is a very specific issue. with its own pathology, requiring great amount of resources because we are talking about state-sponsored activity. this is not about an individual conducting this activity and then you need to review it. so as it relates to
state-sponsored information operations, i'm requesting that you name whoever is responsible now, but as we go forward, that you designate in your operation someone at the executive level who is responsible specifically for those types of operations. understanding that as we know now, there are governments that are willing to put incredible amount of resources into manipulating the american public. and it is beyond what you might need to review in terms of activity on your sites that involves issues of posting inappropriate images and things of that nature. mr. stretch. >> senator, we have a chief security officer and a threat intelligence team that's acutely focused on this threat. i will take responsibility for our overall response to this threat. >> and how many of your employees are dedicated to addressing state-sponsored operations? specifically, and if there aren't, please follow up in terms of what you're prepared to
dedicate to that. >> this is a harder question because there are so many vects that we're investing in. i stated earlier that we have 10,000 people at facebook across a number of teams who are focused on safety and security generally, and we're doubling that number. the number of people who think of this as their full-time job is something i'll have to come back to you on. >> i appreciate that. from each of the companies, we would appreciate that. >> you can create automated systems that detect foreign propaganda. for example, you can determine whether a user is active during moscow business hours or connects through a vpn or registers with a fishy voice over an ip telephone number. and you can feed those signals into a machine that can actually create an algorithm that can allow us to indicate or figure out if propaganda is actually being pushed through. have you done that as it relates to state-sponsored manipulation of elections?
>> so, our technology is agnostic. we have the technology you're talking about, which is an algorithm that helps us catch the bad actors based on their pattern and behavior and also connect accounts so if they start new accounts or new networks of accounts, we get those before they tweet. we want to catch that activity all over twitter. having automated accounts, malicious actors on twitter, is a bad experience for our users. we have been tackling that problem for years. and the challenge is, as we get better, these actors get better. o it's a constant game of cat and mouse and one upsmanship, but we're committed every single day to making sure we're removing those actors from our platform. >> my time is running out. perhaps we can have quick answers for the remaining folks. >> our answer would be similar. >> the same. >> okay, thank you. >> members should be aware that there has been a series of votes started about seven minutes ago. two votes.
we have two members left. we're going to move through those and wrap up. and at this time, i would ask unanimous consent that all members be allowed for seven days to submit questions to our witnesses today. without objection, so ordered. senator cornyn. >> thank you for being here. it strikes me that the united states is operating at a tremendous disadvantage. we're free and open society. we believe in freedom of the press, freedom of expression, and we respect the privacy rights of individuals. our opponents have the opposite view. they view information as a tool of warfare while denying their own citizens access to the sort of freedoms of the press, individual thought and expression, that we celebrate in this country. but it seems to me that you're at the intersection of this problem, which brings us to why you're here today. i'll restate the chairman's
comments when he started by quoting hl minken, who said for every complex problem, there's a clear, simple, and wrong answer. and so we need to be very careful, i think, in how we deal with this. but i do think the public needs to understand how your platforms operate. my friend senator harris talked about you being the town square and the newspaper, the radio station, but you are more than just a publisher of information. as you point out, most of the income that your companies earn is from data mining. you know more about individual americans than anybody else, including their own government. and of course, you vow to protect that information and their privacy at the same time. but you use it in order to target ads. many of us here on the panel have used your platforms and
political campaigns to make sure that our message gets to voters who we think might be receptive to our point of view or our platform. but i would like to know, as a general matter, how do you distinguish between somebody like senator feinstein alluded to using social media as a means to recruit and incite lone wolf terrorists like the one who plowed down unsuspected new yorkers yesterday, killing eight and injuring 11 others. how do you distinguish between that person and a foreign government using your platforms in order to press -- pursue the sort of disinformation and active measures that cause so much confusion and polarization in our election? and how do you distinguish between the way you treat those people on your platforms and how you treat sex traffickers, who are targeting young girls and
selling them in essence for sex. how do you distinguish between those three? >> senator, i would say none of those activities are acceptable on our platforms. we have strong policies against abuse of platforms to promote hate, hate speech, incitement of violence, sex trafficking, human trafficking, et cetera. in addition, when you have the adding layer of a deceptive actor like a foreign government, trying to push its points of view in deceptive ways, without identifying itself, et cetera, that separately violates another layer of our policies. >> talking about what we talked about earlier, we have the same policies prohibiting all of this. we have had the best success so far looking at the behavior and the signals we see only on the twitter side around how accounts are linked and the activity of those accounts, where they're logging in from, what they're doing with each other. to stop these bad actors before they're even able to tweet. a good example of this is on the terrorism side.
we are now able to automatically detect and take down 95% of terrorist accounts. 75% of those before they tweet for the first time. so because we're folking on behavior and not the content, we don't have to wait for them to get their messages out. we take those down proactively before they're able to speak. >> i would just add, senator, that of the three forms of abuse of the platform you describe, child safety, terrorism, and foreign interference in elections, the first two we have been working on for quite a long time as an industry and with government. and i think we have a proven track record of working well together. as an industry and with government. to make sure we're taking appropriate steps to address those abuses. we have work to do, but we have made progress. this threat of foreign interference in the elections is something where we do need to up our game, i think as a company and as an industry, and working with government, but the success we have made or the successes we
have had on those first two forms of abuse gives me some reason for optimism going forward with respect to our ability to address the foreign interference threat. >> for each of you, have the terms of service of your company changed at all since the 2015-2016 actions of russia? your terms of service. >> yes, we have changed our policies with our ads that are appropriate as well as content that's available on youtube. >> focused on active measures or on other matters? >> our terms have to do with terms around advertising and terms about what acceptable content is. we're simultaneously using additional tools behind the scenes to identify material. >> very similar at twitter. >> same is true for facebook. >> why should your companies be treated any different than the press? from a legal accountability
standpoint? >> we believe as a user generated content platform, that the rules around section 230 provide a platform to our users around free speech and expression. and don't require us to take a bias on removing content that we fear will violate certain rights. so we work actively to prohibit things like violence and terrorism and abuse and harassment, and you'll see how we were tackling this problem with urgency and seriousness, but we believe that as a user generated content platform, we want to aplow the free expression and debate without the interference of some of the things you're talking about. >> so you believe you should be treated from a legal standpoint differently than a newspaper, a cable tv show, or a radio show? >> yeah, we're not producing the content. we're allowing users to
upload -- we have a lot of great journalists and news organizations putting content on our platform to share, linking back to their sites. we're offering the service to allow that interchange, that information sharing. >> that may well be a distinction that is lost on most of us. that you're just a platform for other people to express their views as opposed to being a publisher in your own right of those views. finally, let me just ask each of you, please, to continue to work with us on the stop enabling sex traffickers act of 2017. as you know, this deals with the communications decency act, which has been used as a bar, a legal bar to those who have been victims of sexual abuse when they seek to bring their people who facilitated that sex trafficking to justice. and i think there is a way that working together we can come up
with something that protects the victims but also maintains the freedom of the internet and i would just encourage each of you to continue working with us on that so we can reach an acceptable outcome. thank you, mr. chairman. >> senator reed. >> thank you, mr. chairman. to all the panelists, starting with mr. stretch, when you discovered a deceptive foreign government presentation on your platform, my presumption is you said you'll stop it and take it down. do you feel an obligation in turn to notify those people who have accessed that. and can you do that? and shouldn't you do that? >> senator, we feel an obligation, as you say, first to stop the activity. second, to investigate it further. to fan out essentially from the account to make sure we're taking an expansive view of the investigation to try to capture any related activity. third, to share threat information with the industry and with the government so we can all do a better job.
and fourth, to bring the issue to the attention of in this case, this committee. and the content itself, we have said we're supportive of this committee making it publicly available. the question of reaching out to individuals who may have seen it is a much more difficult and complex one, but we believe our commitment to transparency on this issue generally should address that. >> well, technically, you could do that, i presume. or you could invest the resources to do it. as a result, frankly, reporting to us about the nefarious activities of russia is not going to immediately translate to the thousands or apparently 126 million people who saw the message and thought it was legit. you have the, i presume, the technical skill to do that.
and again, you'll see in the newspaper, we correct the statement we made the other day. it was wrong. or it was deliberately wrong. and i think you have given the first amendment, you can live with that, i hope. >> i'm sorry, senator. could you repeat the question? >> well, the question goes back to having an obligation under the first amendment to notify people who you know have been deliberately misled by a foreign government. not just us, not just law enforcement. >> the technical challenges associated with that undertaking are substantial, particularly because much of the data work underneath our estimate of the number of people who may have been exposed to this relies on
data analysis and modeling. that said, we do believe transparency in this area is important. and we are supportive of making as much of this information available to the public as the committee deems warranted. >> i deem it warranted, for what that's worth. mr. edgett? >> it's an interesting pr propositi proposition. we have a team dedicated to quality and how we present information on the platform. we see as an open platform active dialogue around a lot of this false information, fake information, right away. so when you're seeing the tweets, you're also seeing a number of replies to it showing people where to go, where other information is that's accurate. but we will definitely take that idea back to explore how we could implement a process like that. >> mr. walker, your platform? >> we're somewhat differently positioned because we're not primary a social network, many
users are not logged in when they access content, so it's difficult to know who sees what. >> with respect to bots, what i have gleaned from the testimony, and we'll go back to the end, is when you can technically identify a bot system operating on your platform. and then, am i right to assume you will shut them down if -- any bot system you will shut down or particularly a government related bot system? >> when we refer to bots, it's primarily a twitter issue. i'm not familiar to bots on google, per se. >> let's go to -- i'm not the technical expert, but youtube is your subsidiary, i presume. >> that's correct, yes. >> and a lot of the hits on youtube, i presume, were generated by electronic devices, not people. so that the rt program that was
attacking secretary of state clinton, had a 2% audience, but huge number of hits on youtube. and as a result, you know, you are being unwittingly or wittingly used by bot systems all the time. >> so a bot, an automated view of content, really, sort of the core issue, what people try to do on youtube is drive up their perceived view accounts to make themselves appear more popular than they are. >> they do it by using electronic networks. >> yes, and this is a problem not limited to this context. many people would like to make themselves appear more popular than they are, so we have a lot of sophisticated tools that are designed to combat that. >> if you find it, you will reduce the number of hits so it's no longer trending? >> either that or remove people from our services for abuses of
our terms of service. >> you do have to deal with these bot nets? >> in that sense, that's correct. >> bots. >> similarly, we remove as many bad automated accounts as we can find. as i said earlier, some of these bad actors are trained to get more sophisticated. and so we're staying ahead of that by learning from the automated accounts we're seeing. but we will remove them and have technology to make sure that that automated accounts aren't gaming trends so the trending hash tags that people are seeing on the platform, we'll remove their content from search and from the timelines and remove them permanently from the system once we're able to investigate. >> mr. stretch, please. >> senator, apologies. we prohibit automated account creation, and in doing so, we're always looking for evidence of accounts being created, and
engaging in either of the behaviors that mr. walker and mr. edgett identified. >> just a quick question because my time is just about gone. this is a daunting effort. we're being attacked, and you have to go on the offense. the counter offense because the way we have structured this system, we have very limited government role in your regulation, your activities, et cetera. and it comes down ultimately to resources. so i would like you to follow up officially, but what percent of your revenue are you devoting to these activities, proactive activities, not just if someone complains, we'll take it down, but finding bots, thinking about notifying recipients of clearly bad information. so what percent do you think right now? >> senator, i cannot give you a
percentage. i can tell you the company is committed to getting this right. our most expensive resources are people. and as i stated earlier, we are doubling the amount of people who will be focused on these efforts in the coming year. >> can you get us a number please in writing? >> yes. >> thank you. >> mr. edgett, please. >> we also dedicate a lot of resources to this. and i'll follow up with your staff. >> mr. walker. >> similar. >> thank you. >> chair recognizes the vice chair. >> thank you, mr. chairman. i'll be very brief. one, i want to acknowledge senator reid, i think, raised a very good question. if you're in a medical facility and you got exposed to a disease, the medical facility would have to tell the folks they were exposed. the comment also about tv and radio making corrections, it's an interesting question about what obligation you may have. i have to tell you, i think there have been some progress made, and i appreciate some of
o your efforts. i still find it very disturbing that it appears that at least from twitter and facebook, the sense is that all the russian measures only originated with one arm in st. petersburg and it still appears most of the work that you provided us is derivative of your nish reports. i was hoping very much that you would come in today and either say, that was absolutely all or we have identified other troll farms or other entities. i think we have a lot more work to do. thank you so much, mr. chairman, for this hearing. >> thank you, senator warner. we have come to the close of this. and let me just make a couple statements, if i can. if for some reason you need anti-trust waivers to collaborate with each other, please let us know. more importantly, seek for the waiver yourself. this is going to take an overall
effort to minimize. i'm not going to use the term eliminate. to minimize the damage and more importantly the impact of what russia is doing, did do, and what others will do next year. i firmly believe that all three of your companies have a new perspective on security and that you've got -- you're at varying degrees of changes that are good. the challenge is if it fails, the impact of that failure is significantly different than it was in the 2016 elections. i need you to know that up front. you all acknowledge that fec law applies to you. but it hasn't been lost on me that all of you asked for an exemption from the applicable fec law. so, you know, i sort of am
reminded that a portion of the content posted on facebook by foreign actors appeared to support one candidate or another. clearly, it falls within the lines of what the fec law was there for. and as you know, federal campaign law requires disclosures of sources for ads. i have to put that little thing, paid for by. mr. edgett, you said you guys are adopting a new thing. i applaud you on that. and i'm sure if i asked you why you didn't apply it, you might have told me before that the fec law didn't apply to you because they had a hung jury at the fec. they didn't decide it did or didn't. or maybe it's an excuse that it was small, impractical items, therefore, they had an exception to the disclosure. but let me make it perfectly clear. there is no exception to the disclosure as it relates to foreign money used to influence
u.s. elections. that is a national security issue. it is a direct attempt to infiltrate the democracy we have here, and if it's not stated in the law, it should be in your company responsibility to take it on head-on. so in the future, i hope that if there's a takeaway from this, it's that everybody is going to adhere to fec law. if you're a media outlet and you know it applied to you, if for some reason you questioned whether it was foreign money, you probably didn't really act. i hope that none of you's platform is conditioned where you're not verifying where the money is coming from. so the one thing that i didn't hear today and i hope you'll take it back is the effort to certify who's paying for these things. i, like others, do not want the
government to stipulate to any of you what content, especially political content, should look like. by the same token, you're the front line before anything else has to kick in to certify that foreign money is not finding its way into influencing u.s. elections. i wish i could leave today and believe the only thing we have to worry about are elections. the truth is we have spent a tremendous amount of time, two old guys trying to figure out what bots were and things i don't use, but i have to understand them in a way that as a policymaker makes decisions that are best for the future of my kids and my grandchildren. to make sure they've got access to your platforms, to make sure that they can experience things i never dreamed about experiencing and am too old to understand, but i recognize the fact that i can't be influenced just because i don't understand it because i have to match my capital with the intellect of
wrapping up this hearing looking at russia's social media influence on the 2016 election. this is the first of two such meeti ings today. that is scheduled for 2:00 p.m. eastern and we'll have it live for you when it gets started right here on c-span3, and to get us up to 2:00 p.m. eastern and the start of the healouse intelligence hearing, this is a
portion of the meeting that just wrapped up. >> i call the hearing to order. >> good morning, i would like to welcome our witnesses today before i introduce them, i want to say on behalf of the full committee that our hearts and our prayers go out to the individuals of new york, the families and the friends of those who were affected by a senseless terror act. to most on this committee, we have come to expect this. we spend countless hours working through the threats that exist to this country and around the world. and it's sad that we have come to the point where really nothing can happen that surprises us. but it's the responsibility of this committee to work hand in hand with our intelligence community, to help to keep america safe by providing the tools they need to accomplish their mission. we will continue to